Comparison of blue and green water fluxes for different land use classes in a semi-arid cultivated catchment using remote sensing

• Blue and green water fluxes were estimated from four remote sensing methods. Budyko SN methods should be applied with cautions for semiarid catchments. All show similar spatial distribution pattern of blue fluxes. Forested areas Mt. Kilimanjaro Meru consume up to 30 % water. EK method yields the most realistic results. Kikuletwa catchment, Upper Pangani River Basin, Tanzania. This study compared yearly using different methods: Senay’s (SN) (Senay et al., 2016), van Eekelen (EK) (van 2015), (Simons 2020) Soil Water Balance (SWB) model (FAO IHE Delft, 2019). The Land Use Cover (LULC) classes an ensemble seven global sensing-based evapotranspiration products (Ensemble ET) CHIRPS rainfall dataset. Ensemble ET was created RS-based surface energy balance models (GLEAM, CMRS-ET, SSEBop, ALEXI, SEBS, ETMonitor MOD16). Our found that able map results irrigated non-irrigation cultivated areas. SWB gave too high non-irrigated agricultural areas, whereas not a clear difference in blue-water versus On other hand, no more than half identified Three estimate highest (318–582 mm/y) forested while estimates banana coffee (278 mm/y). Overall, we conclude yielded when land use map, could considered after further calibration if higher temporal data resolution (e.g. monthly) is required.